Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: -1 observations deleted


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: -1 observations deleted
Date   Wed, 14 Aug 2002 15:00:16 +0100

"Lee Chuntao" <chtl@econ.hku.hk>

>       I need to connect a lot of files together by using the command
"append
> using". Those files all contain the same six variables, but
different
> numbers of observations. Some contain 0 observations. I also need to
delete
> some observations in some conditions. For the zero-observation
files, surely
> i deleted zero observation by using "drop if condition", but stata
replied
> "-1 observations deleted". This is something easily be
misunderstood. My
> question is, why stata not report "0 observation deleted".

I agree that's a puzzling message.

But I don't understand what the real problem is.

If you don't know that a file contains 0 observations
you can check ahead of reading it in:

qui d using file
if r(N) > 0 {
	append using file
}

That is, a -describe- without reading in the file
will leave behind a count of the number of observations,
which you can then use to avoid reading in an empty data set.

>       another question is also related to zero-observation files.
there is a
> variable named date, and i need to sort obs. decendingly. If i use
> "gsort -date", there is an error, but if i use the following
commands, it
> works well:

>                gen n_date = -date
>                gsort n_date
>                drop n_date

> this seems rather  stupid since we have a command "gsort -varname"
> can stata co. upgrade the gsort command?

Let me get this straight. You are reporting a bug: you
cannot use -gsort- to reverse sort a file with 0
observations. OK. You could get round this by
writing your own copy of -gsort-, say -gsort2-,
which would be the same as -gsort- except that
early on you have these lines

qui count
if r(N) < 2 {
	exit 0
}

which will throw you out gently without
changing your data set (the way I see it,
any data set with <2 observations is invariant under any kind of
-sort-, so the -sort- is redundant).

If the problem is that you are doing this
in an automated way, and you are not
clear in advance which files have
zero observations, then my answer is as above.

Nick
n.j.cox@durham.ac.uk

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index